Effects of Cooling Rate on the Solidification and Microstructure Of

materials

Article Eﬀects of Cooling Rate on the Solidiﬁcation and Microstructure of Nickel-Based Superalloy GTD222

1 1 1 1 1 2, , Bo Gao , Yanfei Sui , Hongwei Wang , Chunming Zou , Zunjie Wei , Rui Wang * † 2,3, , and Yanle Sun * † 1 National Key Laboratory for Precision Hot Processing of Metals, School of Materials Science and Engineering, Harbin Institute of Technology, Harbin 150001, China; [email protected] (B.G.); [email protected] (Y.S.); [email protected] (H.W.); [email protected] (C.Z.); [email protected] (Z.W.) 2 Shanghai Key Laboratory of Advanced High-Temperature Materials and Precision Forming, School of Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, China 3 School of Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, China * Correspondence: [email protected] (R.W.); [email protected] (Y.S.) These authors contributed equally to this work. † Received: 6 May 2019; Accepted: 12 June 2019; Published: 14 June 2019

Abstract: In this work, the microstructure and solidification behavior of nickel-based superalloy GTD222 at different cooling rates are studied. The solidification of the superalloy GTD222 proceeds as follows: L L + γ,L L + γ + MC, L L + (γ/γ )-Eutectic and L η phase. Due to alloying element → → → 0 → redistribution, the temperature of the solidus GTD222 superalloy, 1310 ◦C, is slightly lower than the temperature of the liquidus, which is 1360 ◦C. It was found that the dendrite arm spacing of the alloy decreased with the increase of the cooling rate from 200 µm at 2.5 K/min to 100 µm at 20 K/min.

Keywords: GTD222; nickel-based superalloy; solidiﬁcation behavior; cooling rate

1. Introduction Casting superalloys are widely applied in industrial areas such as gas turbines, aerospace and chemical process industries owing to their excellent mechanical properties and thermal corrosion resistance [1,2]. In order to develop more efficient advanced solidification technology, a data base of superalloys on thermophysical properties is increasingly needed [3]. From a theoretical and industrial perspective, knowledge of superalloy solidification behavior is crucial for the controlling of the superalloy casting process [4]. In the casting of nickel-based superalloys, the mechanical properties of alloys are due to microstructure characteristics, such as a combined matrix γ phase and γ0 precipitation phase [5], dendritic width [6] and grain size [7]. Hence, the optimum mechanical properties can be obtained only by applying suitable heat treatment [3]. For heat treatment after casting, the solidus temperature (the temperature at which incipient alloys begin to melt) and the formation temperature of precipitation (the onset precipitating of the γ0 phase, η phase and carbide phase), which determine the heat treatment window of materials, are very important. To obtain the ideal microstructure, heat treatment of the nickel-based superalloys must be carried out in certain temperature range. Therefore, for any nickel-based superalloy, solvus temperature and solidus temperature should be accurately measured, in order to optimize the mechanical properties of the superalloy. What is more, the incipient dissolving temperature of precipitation is another important parameter, helping to maximize the volume of precipitates without producing an interdendritic region. In recent years, the demand of hot-end complex parts with different wall thicknesses has increased. The casting system of the hot-end parts is very intricate, and that leads to difficulty in controlling

Materials 2019, 12, 1920; doi:10.3390/ma12121920 www.mdpi.com/journal/materials Materials 2019, 12, 1920 2 of 9 the microstructure. Numerous studies have shown a direct relationship between the cooling rate and material microstructure, such as the significant effect on the dendritic width by the cooling rate of solidification process. Chen et al. [8] studied the compositional changes of the micro-scale precipitates of an advanced Ni-base superalloy at different cooling rates. It was found that the chemical composition of the precipitates of different sizes is very different. This study has important implications for understanding the microstructure and precipitation behavior of Ni-based superalloys. Zheng et al. have testified that the cooling rate significantly influences the morphology of dendrites [9]. The dendrite arm spacing of both primary and secondary dendrite declined as the cooling rate increased [10–12]. GTD222, a nickel-based precipitation hardened isometric crystal superalloy, is considered to be one of the most suitable superalloys that can be processed into the guide vane of a steam turbine, servicing at 1000 ◦C[13]. Most work focused on the optimization of the composition of the GTD222 superalloy. However, less attention has focused on the solidification behavior of the GTD222 superalloy [14]. In this study, an empirical research was carried out to understand the effect of cooling rate on the solidification behavior and microstructural evolution of the GTD222 superalloy, and the liquidus temperature and solidus temperature of the GTD222 superalloy were also measured.

2. Experimental Procedures The chemical composition of the GTD222 superalloy used in the present work is given in Table1. Commercial pure metals (> 99.95 wt.%) were used for the preparation of the alloy prior to melting. To ensure compositional homogeneity, the alloy melt was fully stirred by an electromagnetic stirring system equipped in the arc furnace (QSH-ZP, Quanshuo, Shanghai, China), and each button alloy was ﬂipped and melted at least four times. The materials were ﬁrstly prepared in a vacuum induction melting furnace, and then casted into ingots (100 mm 100 mm 150 mm). All specimens used in this × × work were cut from the ingot using a spark cutting machine (Dk77, Zhonggu, Suzhou, China).

Table 1. Chemistry (wt.%) of the GTD222 superalloy.

C Cr Co W Al Ti Nb B Ta Ni 0.08–0.12 22.2–22.8 18.5–19.5 1.8–2.2 1.0–1.4 2.1–2.5 0.7–0.9 0.002–0.007 0.9–1.1 Bal.

The solidification procedure of the GTD222 superalloy was revealed by differential scanning calorimetry (DSC, DSC 404 F3 Pegasus, NETCH, Selb, Germany). All DSC testes were carried out by alumina crucibles in the argon protection environment with sample sizes of 2.5 mm in diameter and 2 mm in height. The cooling rates of samples were 2.5 K/min, 5 K/min, 10 K/min and 20 K/min, respectively. Microstructural evolution and phase constitutions of GTD222 samples were carried out on an optical microscope (OM, Axio Imager A1m, ZEISS, Jena, Germany), X-raydiffraction (XRD, XRD-6000 diffraction instrument, Shimadzu, Kyoto, Japan), X-ray energy-dispersive spectroscopy (EDS, JSM-7600F, Tyoto, Japan) and scanning electronmicroscope (SEM, Sirion 200, FEI, Hillsboro, OR, USA). The samples for OM and SEM were ground to 2000 grit, and then polished by the diamond polishing paste (1 µm). The etchant for samples was 45 mL CuSO4, 100 mL H2O and 50 HCl. Phase constitutions of the alloy were determined by X-ray diffraction (XRD) technique, using Cu Kα (λ = 0.1540562 nm) radiation, operating at 40 kV and 30 mA between 20◦ and 80◦ (2θ) at a step of 0.02◦ and a counting time of 0.6 s per step.

3. Results and Discussion

3.1. Microstructure of As-Cast Alloy Figures1 and2 show the microstructure of an as-cast GTD222 superalloy with di ﬀerent cooling rates. The microstructure of the GTD222 superalloy is a typical dendritic structure in the as-casting samples, and no equiaxed grains were found in Figure1. The dendritic structures in Figure1a are coarser than dendritic structures in Figure1b–d. It can be seen that the dendritic structures are Materials 2019, 12, x FOR PEER REVIEW 3 of 9

Figures 1 and 2 show the microstructure of an as-cast GTD222 superalloy with different cooling rates. The microstructure of the GTD222 superalloy is a typical dendritic structure in the as-casting Materialssamples2019, and, 12 ,no 1920 equiaxed grains were found in Figure 1. The dendritic structures in Figure 1a3 are of 9 coarser than dendritic structures in Figure 1b–d. It can be seen that the dendritic structures are coarsening, along with the decrease in the cooling rate. Additionally, the secondary dendritic arm coarsening,spacing has alongthe same with tendency. the decrease According in the cooling to V. Kavoosi, rate. Additionally, the secondary the secondarydendritic arm dendritic spacing arm is spacingrelated has to the the same local tendency. cooling rate According [15]. The to V. size Kavoosi, of the the secondary secondary dendritic dendrites arm directly spacing affects is related the tocomposition the local cooling segregation, rate [ 15the]. sec Theond size phase of theand secondary the distribution dendrites of micropores. directly aff Asects shown the composition in Figure 2, segregation,the size of the the carbide second precipitation phase and the is distribution decreasing ofas micropores.the cooling Asrate shown increas ines Figure from2 2.5, the K/min size of to the 20 carbideK/min, since precipitation a low cooling is decreasing rate would as the provide cooling more rate increases time for fromthe diffusion 2.5 K/min of to the 20 atom K/min,s and since coarsen a low coolingthe second rate phase. would Furthermore, provide more the time secondary for the didendriteffusion arm of the spacing atoms of and the coarsen alloy is theclosely second related phase. to Furthermore,tensile strength the and secondary elongation. dendrite Generally, arm spacing the smaller of the the alloy distance is closely between related secondary to tensile dendritic strength andarms, elongation. the better the Generally, mechanical the smallerproperties the p distanceroduced. between The high secondary cooling rate dendritic would arms,refine thethe bettergrain theand mechanical dendrite arm properties spacing of produced. materials, The producing high cooling a greater rate wouldfraction refine of boundaries the grain and dendriteconsequently arm spacingimproving of materials, the mechanical producing performance a greater of fraction the materials of boundaries at ambient and consequently temperature. improving However, the mechanicalboundaries performancewould become of weak the materials phase in atthe ambient elevated temperature. temperature However, atmosphere the [16, boundaries17]. More would grain becomeboundaries weak and phase finer in dendritic the elevated arm temperature spacing would atmosphere make the [16 mechanical,17]. More properties grain boundaries of the materials and finer dendriticdecrease armdramatically. spacing would Therefore, make it the is mechanical important properties to find a suita of theble materials cooling decrease rate to optimize dramatically. the Therefore,potential of it isthe important material to to find control a suitable the microstructure cooling rate to of optimize the material the potential and thus of to the optimize material the to controlproperties the of microstructure the superalloy. of the material and thus to optimize the properties of the superalloy.

Figure 1.1. Optical microscope ( (OM)OM) images of the as-castas-cast GTD222 superalloy with didifferentﬀerent coolingcooling ratesrates during the the solification soliﬁcation process, process, (a (a) )cooling cooling at at 2.5 2.5 K/min, K/min, (b ()b cooling) cooling at at5 K/min, 5 K/min, (c) (coolingc) cooling at 10 at 10K/min, K/min, (d) ( dcooling) cooling at at20 20K/min. K/min.

Materials 2019, 12, 1920 4 of 9 Materials 2019, 12, x FOR PEER REVIEW 4 of 9

Figure 2. SEMSEM images images of of the the as as-cast-cast GTD222 GTD222 superalloy superalloy with with different different cocoolingoling rateratess during the solificationsolification process, (a) cooling at 2.5 K/min, K/min, ( b) cooling at 5 K/min, K/min, ( c) cooling cooling at at 10 10 K/min, K/min, ( d)) cooling cooling at 20 K/min. K/min. 3.2. Solidification Process 3.2. Solidification Process Figure3 shows the heating DSC curves of the GTD222 superalloy sample cutting from the Figure 3 shows the heating DSC curves of the GTD222 superalloy sample cutting from the asas-casting ingot, the heating rate of the sample is 2.5 K/min. The solidus temperature of the matrix casting ingot, the heating rate of the sample is 2.5 K/min. The solidus temperature of the matrix γ γ phase can be identified from the heating curve. The value of γ phase solidus temperature is phase can be identified from the heating curve. The value of γ phase solidus temperature is 1310 °C. 1310 C. The ending temperature of the endothermic peak corresponding to the matrix γ phase is The ending◦ temperature of the endothermic peak corresponding to the matrix γ phase is 1380 °C. 1380 C. According to the DSC curve, the beginning dissolution temperature of the γ phase is 1190 C. According◦ to the DSC curve, the beginning dissolution temperature of the γ′ phase0 is 1190 °C. The◦ The temperatures of 1240 C and 1325 C represent the ending solution temperature of γ phase and temperatures of 1240 °C and◦ 1325 °C represent◦ the ending solution temperature of γ′ phase0 and the the beginning solution temperature of the carbides, respectively. It is worth noting that it is difficult to beginning solution temperature of the carbides, respectively. It is worth noting that it is difficult to retrieve the dissolving values of γ phase from the measurements of the DSC heating curve, owing to retrieve the dissolving values of γ′0 phase from the measurements of the DSC heating curve, owing to the noisy background accompanied with sluggish heating rate of 2.5 K/min [3]. the noisy background accompanied with sluggish heating rate of 2.5 K/min [3]. Figure4 shows the DSC curves of samples with a cooling rate of 2.5 K /min, 5 K/min, 10 K/min and Figure 4 shows the DSC curves of samples with a cooling rate of 2.5 K/min, 5 K/min, 10 K/min 20 K/min, respectively. All the DSC samples were heated to 1500 C with a heating rate of 2.5 K/min and 20 K/min, respectively. All the DSC samples were heated to◦ 1500 °C with a heating rate of 2.5 and held for 10 min at this temperature prior to the cooling process down to 700 C. Taking the curve K/min and held for 10 minutes at this temperature prior to the cooling process down◦ to 700 °C. Taking of the sample with a cooling rate of 20 K/min as an example, two obviously exothermic peaks were the curve of the sample with a cooling rate of 20 K/min as an example, two obviously exothermic observed as the sample cooled down to 700 C. From high temperature to low temperature, the first peaks were observed as the sample cooled◦ down to 700 °C. From high temperature to low and largest exothermic peak is related to the process of solidification of the matrix γ phase, which can temperature, the first and largest exothermic peak is related to the process of solidification of the be described as the following formula: L-L + γ. The initiating temperature of this exothermic peak of matrix γ phase, which can be described as the following formula: L-L + γ. The initiating temperature the cooling DSC curve is 1360 C, closer to the other three curves with different cooling rates. The end of this exothermic peak of the◦ cooling DSC curve is 1360 °C, closer to the other three curves with temperature of this exothermic peak is about 1293 C, slightly lower than the other curves. It can different cooling rates. The end temperature of this exothermic◦ peak is about 1293 °C, slightly lower be seen from the curves that the ending temperature of the first peak decreased as the cooling rate than the other curves. It can be seen from the curves that the ending temperature of the first peak increasing from 2.5 K/min to 20 K/min. Followed by the first exothermic peak, the second exothermic decreased as the cooling rate increasing from 2.5 K/min to 20 K/min. Followed by the first exothermic peak with an ending temperature of 1281 C was observed. The heating curve was used to obtain peak, the second exothermic peak with an ending◦ temperature of 1281 °C was observed. The heating the solidus, and the cooling curve was used to quote the liquidus [3]. Therefore, from the curves curve was used to obtain the solidus, and the cooling curve was used to quote the liquidus [3]. shown in Figures3 and4, the solidus of the GTD222 superalloy can be determined as 1310 C and the Therefore, from the curves shown in Figures 3 and 4, the solidus of the GTD222 superalloy◦ can be liquidus can be determined as 1360 C. Figure4 also shows that the exothermic peak of matrix γ phase determined as 1310 °C and the liquidus◦ can be determined as 1360 °C. Figure 4 also shows that the became sharper when the cooling rate increased from 2.5 K/min to 20 K/min, with a slight decline of the exothermic peak of matrix γ phase became sharper when the cooling rate increased from 2.5 K/min non-equilibrium phase transition temperature. Since undercooling of materials increased along with to 20 K/min, with a slight decline of the non-equilibrium phase transition temperature. Since the increasing of cooling rate, the enhanced undercooling of the sample with a high cooling rate would undercooling of materials increased along with the increasing of cooling rate, the enhanced rapidly release large latent heat, leading to the recalescence effect that remelted the solidified primary γ undercooling of the sample with a high cooling rate would rapidly release large latent heat, leading phase. The partial re-melted matrix γ phase would cause solidification and release the heat again [18]. to the recalescence effect that remelted the solidified primary γ phase. The partial re-melted matrix γ That is why the exothermic peak of γ phase is becoming sharper and the ending temperature of the first phase would cause solidification and release the heat again [18]. That is why the exothermic peak of γ phase is becoming sharper and the ending temperature of the first peak becoming lower as the

MaterialsMaterials 20192019,, 1212,, x 1920x FORFOR PEERPEER REVIEWREVIEW 55 of of 9 9 coolingcooling rate rate increases. increases. AA l largearge ccoolingooling rate rate from from the the molten molten state state would would produce produce an an enhanced enhanced undercooling,peakundercooling, becoming increaseincrease lower as thethe the drividrivi coolingngng forceforce rate ofof increases. thethe formationformation A large ofof coolingnonnon--equilibriumequilibrium rate from phasephase the molten,, andand promotepromote state would thethe precipitationproduceprecipitation an enhanced ofof thethe nonnon undercooling,--equilibriumequilibrium phase.phase. increase ForFor the thethe driving competcompet forceiitiontion of betweenbetween the formation thethe nonnon of--equilibriumequilibrium non-equilibrium phasephase andphase,and thethe and primaryprimary promote γγ phase,phase, the precipitation dendritesdendrites atat of largerlarger the non-equilibrium coolingcooling ratesrates areare phase. finerfiner For thanthan the thethe competition dendritesdendrites between obobtainedtained the atat smallnon-equilibriumsmall coolingcooling rates,rates, phase asas shownshown and the inin primaryFigureFigure 1.1.γ phase, dendrites at larger cooling rates are finer than the dendritesOOnn accountaccount obtained ofof atthethe small complexitycomplexity cooling ofof rates, thethe phase asphase shownss inin the inthe Figure GTD222GTD2221. superalloy,superalloy, itit isis arbitraryarbitrary toto identifyidentify whatwhatOn thethe account exothermicexothermic of the peakpeak complexity isis toto referrefer of toto the thethe phases curvecurve itself.itself. in the Therefore,Therefore, GTD222 superalloy, thethe phasephase analysisanalysis it is arbitrary,, andand accordinglyaccordingly to identify twhatthehe solidificationsolidification the exothermic processprocess peak ofof is thethe to refer GTD222GTD222 to the superalloysuperalloy curve itself.,, waswas Therefore, carriedcarried outout the usingusing phase XRD.XRD. analysis, FigureFigure and 55accordingly isis thethe XRDXRD patternthepattern solidification of of thethe GTD222GTD222 process ofsuperalloy superalloy the GTD222 obtained obtained superalloy, with with wasaa coolingcooling carried rateout rate using of of 2.5 2.5 XRD. K/min K/min Figure and and5 is 20 20 the K/min, K/min, XRD respectively.patternrespectively. of the FigureFigure GTD222 55 showsshows superalloy thethe phasesphases obtained ofof thethe with GTD222GTD222 a cooling superalloy,superalloy, rate of 2.5 includingincluding K/min and γ′γ′ 20phase,phase, K/min, MCMC respectively. phasephase andand ηFigureη phase.phase.5 showsFigFigureuress the 66––88 phases showshow differentdifferent of the GTD222 phasesphases observedobserved superalloy, inin the includingthe interdendritinterdendritγ0 phase,icic areasareas MC ofof the phasethe samples.samples. and η FigureFigurephase. 6Figures6 isis aa typicaltypical6–8 show morphologymorphology di fferent ofof phases thethe γγ observed ++ γ′γ′ eutecticeutectic in the phasephase interdendritic ofof thethe GTD222GTD222 areas supersuper of thealloyalloy samples. thatthat waswas Figure denselydensely6 is a distributedtypicaldistributed morphology inin thethe interdendritic.interdendritic. of the γ + γ0 eutecticFigureFigure 77 phase showsshows of thethe the carbidecarbide GTD222 phasephase superalloy precipitation.precipitation. that was AsAs densely shownshown distributed inin FigureFigure 7bin7b, the, thethe interdendritic. carbidecarbide phasephase Figure precipitationprecipitation7 shows is theis richrich carbide inin Ti,Ti, phase NbNb andand precipitation. Ta.Ta. FigureFigure 8 As8 isis shown thethe typicaltypical in Figure morphologymorphology7b, the carbide ofof thethe ηphaseη phasephase precipitation withwith aa coarsecoarse is needlneedl rich inee--like Ti,like Nb configurationconfiguration and Ta. Figure reportedreported8 is the inin typical castcast NiNi morphology--basedbased alloyalloy [19,[19, of the20]20].. ηNoteNotephase thatthat with onlyonly a thecoarsethe γ′γ′ phasephase needle-like cancan bebe configuration observedobserved inin Figure reportedFigure 66,, insincesince cast thethe Ni-based γγ phasephase alloy waswas [ 19tootoo,20 fragilefragile]. Note toto that resistresist only thethe the corrodentcorrodentγ0 phase andand can disappearedbedisappeared observedin duringduring Figure dd6etectingetecting, since the.. TTheheγ phase morphologymorphology was too ofof fragile thethe γ′γ′ to phasephase resist alsoalso the corrodent variedvaried whenwhen and thethe disappeared Ti/AlTi/Al atomatom during ratioratio changeddetecting.changed [21][21] The.. TT morphologyhehe γ′γ′ phasephase inin of γγ the//γ′γ′ γeutecticeutectic0 phase phasephase also varied isis needleneedle when likelike the whilewhile Ti/Al thethe atom primaryprimary ratio changed γ′γ′ phasephase [ isis21 cubecube]. The likelikeγ0 inphasein thethe present inpresentγ/γ0 eutectic work.work. phase is needle like while the primary γ0 phase is cube like in the present work.

FigureFigure 3.3. HeatingHeating di differentialdifferentialﬀerential scanning scanningscanning calorimetry calorimetrycalorimetry (DSC) ((DSCDSC curve)) curvecurve of the ofof the as-castthe asas--castcast GTD222 GTDGTD superalloy222222 superalloysuperalloy with withawith heating aa heatingheating rate of raterate 2.5 ofof K /2.52.5min. K/min.K/min.

FigureFigure 4.4. DSCDSCDSC curves curvescurves of ofof the thethe alloy alloyalloy cooled cooledcooled from fromfrom super-solidus supersuper--solidussolidus temperature, temperature,temperature, the cooling thethe coolingcooling rates are ratesrates 2.5 areare K/min, 2.52.5 K/min,5K/min, K/min, 55 K/min, 10K/min, K/min 1010 and K/minK/min 20 K andand/min, 2020 respectively.K/min,K/min, respectively.respectively.

Materials 2019, 12, 1920 6 of 9 Materials 2019, 12, x FOR PEER REVIEW 6 of 9 Materials 2019, 12, x FOR PEER REVIEW 6 of 9

FigureFigure 5. The 5.XRDThe patterns XRD of patterns the GTD222 of the superalloy GTD222 with superalloy a cooling with rate aof cooling2.5K/min rate and of 20 2.5K/min, K/min respectively. and Figure 5.20 The K/ min,XRD respectively.patterns of the GTD222 superalloy with a cooling rate of 2.5K/min and 20 K/min, respectively.

Figure 6. Petal γ/γ′ eutectic phases in the interdendritic area of the DSC sample with a cooling rate of 20 K/min. Figure 6.Figure Petal γ/γ′ 6. Petal eutecticγ/γ0 phaseseutectic in phases the interdendritic in the interdendritic area of the area DSC of thesample DSC with sample a cooling with a coolingrate of 20 rate K/min. of 4. Discussion20 K/min. 4. Discussion 4.1.4. Discussion Formation of the γ/γ′ Eutectic Phases 4.1. Formation of the γ/γ′ Eutectic Phases 4.1. FormationThe γ/γ′ eutectic of the γ/ γphase0 Eutectic is not Phases thermodynamic equilibrium, and the emergence of this phase can be T attributehe γ/γ′ eutecticd to the phase solution is not segregation thermodynamic of elements equilibrium, that occurr and thes during emergence the solidification of this phase ofcan the The γ/γ0 eutectic phase is not thermodynamic equilibrium, and the emergence of this phase be GTD222 attribute superalloyd to the solution [22]. The segregation segregation of behavior elements of that the elementsoccurrs during in alloys the is solidificationinevitable, since of thethere can be attributed to the solution segregation of elements that occurrs during the solidification of the GTD222are differences superalloy in[22] the. The diffusion segregation rate behaviors and melting of the elements points of in alloys different is inevitable, elements. since During there the GTD222 superalloy [22]. The segregation behavior of the elements in alloys is inevitable, since there are aresolidification differences of in the the GTD222 diffusion superalloy, rates and the Al melting is rejected point intos of the different liquid phase elements. continuously. During At the last, differences in the diffusion rates and melting points of different elements. During the solidification of solidificationthe content of of the Al GTD222 with a lowersuperalloy, melting the pointAl is rejected is rich enoughinto the toliquid form phase the γ continuously./γ′ eutectic phase At last,. The the GTD222 superalloy, the Al is rejected into the liquid phase continuously. At last, the content of Al theformation content ofprocess Al with of γa/ γ′lower eutectic melting phase point of GTD222 is rich could enough be depicted to form asthe following: γ/γ′ eutectic the first phase emerging. The with a lower melting point is rich enough to form the γ/γ0 eutectic phase. The formation process of γ/γ0 formationphase γ processis precipitated of γ/γ′ eutecticin the liquid phase phase of GTD222 while the could Al inbe the depicted liquid asis rejecfollowing:ted into the the first residual emerging liquid eutectic phase of GTD222 could be depicted as following: the first emerging phase γ is precipitated in phasephase; γ is with precipitated the increas in thee of liquid Al content phase whilethe γ /theγ′ eutectic Al in the phase liquid, Tiis rejecis continuouslyted into the residualrejected liquidinto the the liquid phase while the Al in the liquid is rejected into the residual liquid phase; with the increase of phase;liquid with phase. the Theincreas formatione of Al ofcontent the γ/ γ′the eutectic γ/γ′ eutectic phases phase and γ′, Ti phase is continuously during the solidificationrejected into ofthe the Al content the γ/γ0 eutectic phase, Ti is continuously rejected into the liquid phase. The formation liquidGTD222 phase. superalloy The formation should of occur the γ /almostγ′ eutectic simultaneously, phases and γ′and phase it is during hard to the separate solidification the exothermic of the of the γ/γ0 eutectic phases and γ0 phase during the solidification of the GTD222 superalloy should GTD222peaks relatesuperalloyd to those should two occur phases almost completely. simultaneously, and it is hard to separate the exothermic occur almost simultaneously, and it is hard to separate the exothermic peaks related to those two peaks relateThe thirdd to those phase two found phases in the completely. as-cast samples of the GTD222 superalloy is carbide precipitation. phases completely. FigureThe third7a is phasethe typical found morphology in the as-cast of samples the carbide of the precipitation GTD222 superalloy phase, which is carbide can beprecipitation. identified by FigureEDS, 7a giving is the the typical elements morphology the composition of the carbide that was precipitation shown in Figure phase, 7b .which The elements can be identifiedcomposition by in EDS,Figure giving 7b theis typical elements of the the components composition of that carbide, was shown which inis Figurerich in 7bTi,. NbThe and elements Ta elements. composition The molar in Figureratio 7b of isthe typical metal ofelement the components (the sum of of Ti, carbide, Nb and which Ta) to is the rich carbon in Ti, elementNb and Tais about elements. 1: 1. TheBased molar on the ratio of the metal element (the sum of Ti, Nb and Ta) to the carbon element is about 1: 1. Based on the

Materials 2019, 12, 1920 7 of 9

The third phase found in the as-cast samples of the GTD222 superalloy is carbide precipitation. Figure7a is the typical morphology of the carbide precipitation phase, which can be identiﬁed by

EDS,Materials giving 2019, 12 the, x elementsFOR PEER REVIEW the composition that was shown in Figure7b. The elements composition7 of in 9 Figure7b is typical of the components of carbide, which is rich in Ti, Nb and Ta elements. The molar Materials 2019, 12, x FOR PEER REVIEW 7 of 9 ratiooutcomes of the of metal EDS, elementthe carbide (the can sum be of the Ti, Nbsame and as Ta)the toMC the-type carbon carbide element [23]. is It about is hard 1:1. to Based find out on the outcomesexothermic of peak EDS, of the carbide carbide on can the becooling the same curve as of the DSC MC-typeMC but-type the carbide endothermic [23] [23].. It It peakis is hard of to carbide find ﬁnd out on the exothermicheating curve peak of DSCof carbide can be on easily the coolingconfirmed. curve of DSC but the endothermic peak of carbide on the heating curve of DSC can be easily conﬁrmed.confirmed.

Figure 7. Carbide phase in the solidified DSC sample with cooling rate of 20 K/min. (a) is the SEM Figuremorphology 7. CarbideCarbide of carbide. phase phase ( inb in) theis the solidified EDS spectrum DSC sample of the with spot of cooling carbide rate in ofthe 20 DSC KK/min./min. sample ( a) ,is according the SEM morphologyto the white ofcrossof carbide.carbide. in blue ( b(b) background) is is the the EDS EDS spectrum. spectrumThe inset of theofin thetop spot spotright of carbide of corner carbide in is the thein DSC thecomponent DSC sample, sample accordingratio, accordingin weight to the whitetoand the atom. crosswhite in cross blue in background. blue background The inset. The in topinset right in top corner right is corner the component is the component ratio in weight ratio andin weight atom. and atom.

FigureFigure 8. η 8. phaseη phase in inthe the interdendritic interdendritic area area of of the the solidified solidified DSC DSC samples samples with with cooling cooling rate rate of 20 K/min K/min.. 4.2.Figure Formation 8. η phase of the inη thePhase interdendritic area of the solidified DSC samples with cooling rate of 20 K/min. 4.2. Formation of the η Phase Some researchers have confirmed that the ultimate formation phase of nickel-based superalloy 4.2. FormationSome researchers of the η Phase have confirmed that the ultimate formation phase of nickel-based superalloy with B element is the boride phase, which is rich in Mo and Cr elements [24]. However, in the present with B element is the boride phase, which is rich in Mo and Cr elements [24]. However, in the present studySome no Mo researchers or Cr rich have phase confirmed was found that and the no ultimate exothermic formation peak corresponding phase of nickel to- boridebased superalloy was found study no Mo or Cr rich phase was found and no exothermic peak corresponding to boride was found withon the B element DTA curve is the of boride the GTD222 phase, superalloy. which is rich The in Mo distribution and Cr elements mapping [24] of. theHowever, B element in the is present hard to on the DTA curve of the GTD222 superalloy. The distribution mapping of the B element is hard to studytrace owningno Mo or to Cr its rich low phase content. was Infound the presentand no exothermic superalloy, peak the formation corresponding of η phase to boride represents was found the trace owning to its low content. In the present superalloy, the formation of η phase represents the onending the DTA of the curve solidification of the GTD222 process superalloy. of the GTD222 The superalloy.distribution During mapping the of subsequent the B element cooling, is hard the γto ending of the solidification process of the GTD222 superalloy. During the subsequent cooling, the γ′0 precipitatestrace owning start to its to precipitatelow content. from In thethe matrixpresentγ superalloy,phase, which the corresponds formation toof theη phase third represen inconspicuousts the precipitates start to precipitate from the matrix γ phase, which corresponds to the third endexothermicing of the peak solidification on the cooling process DSC of curve. the GTD222 The onset superalloy. temperature During of the the third subsequent exothermic cooling, peak is the hard γ′ precipitatesinconspicuous start exothermic to precipitate peak on from the thecooling matrix DSC γ curve. phase, Th whiche onset corresponds temperature to of the the third to determine. This is due to the difference in elements solute distribution between dendrite area and inconspicuousexothermic peak exothermic is hard to peakdetermine. on the This cooling is due DSC to the curve. difference The onset in elements temperature solute ofdistribution the third interdendritic area. As a result, the precipitation temperature of the γ0 phase is different in different exothermicbetween dendrite peak is area hard and to determine.interdendritic This area. is due As toa result,the difference the precipitation in elements temperature solute distribution of the γ′ areas. The stack of the exothermic heat of all the γ0 phases in the different regions forms the third betweenphase is dendrite different area in d ifferentand interdendritic areas. The area. stack As of thea result, exothermic the precipitation heat of all temperature the γ′ phases of inthe the γ′ exothermic peak on the cooling DSC curve and that is why the third exothermic peak temperature phasedifferent is regions different forms in d ifferentthe third areas. exothermic The stack peak of on the the exothermiccooling DSC heat curve of and all thethat γ′ is phaseswhy the in third the range is so large. differentexothermic regions peak formstemperature the third range exothermic is so large. peak on the cooling DSC curve and that is why the third exothermicBased onpeak the temperature discussion above, range isthe so formation large. sequence of phase during the solidification of the GTD222Based superalloy on the discussion can be summarized above, the as formation follows: liquidsequence translate of phases into during liquid the and solidification the γ phase, liquidof the GTD222translate superalloys into liquid can and be the summarized carbide phase, as follows: liquid liquid translate translates intos liquid into liquid and the and γ the/γ′ eutecticγ phase, phaseliquid, translateliquid translates into liquids into andthe ηthe phase. carbide The phase, onset liquidtemperature translates ofs intoprecipitation liquid and for the all γ /theseγ′ eutectic phase phases vary, liquidwhen thetranslate coolings into rate the of alloyη phase. is changed. The onset When temperature the coolings of rate precipitation is 20 K/min, for the all onsetthese temperaturephases vary when the cooling rate of alloy is changed. When the cooling rate is 20 K/min, the onset temperature

Materials 2019, 12, 1920 8 of 9

Based on the discussion above, the formation sequence of phase during the solidification of the GTD222 superalloy can be summarized as follows: liquid translates into liquid and the γ phase, liquid translates into liquid and the carbide phase, liquid translates into liquid and the γ/γ0 eutectic phase, liquid translates into the η phase. The onset temperatures of precipitation for all these phases vary when the cooling rate of alloy is changed. When the cooling rate is 20 K/min, the onset temperature of the second exothermic peak is 1293 ◦C. Compared with a sample cooling at 10 K/min, it can be observed that the solidus of a 20 K/min sample is slightly lower than the solidus of a sample cooling at 10 K/min. The phase transformation of the GTD222 superalloy solidification behavior under different cooling rates, the solidus temperature, and the liquidus temperature are important components of GTD222 superalloy thermophysical data. Those thermophysical data are the basis for designing the casting process of the alloy and exploring the potential capacity of the superalloy by using heat treatment.

5. Conclusions 1. The solidification of the GTD222 superalloy proceeds as follows: L L + γ,L L + γ + MC, → → L L + (γ/γ )-Eutectic and L η phase. The type of carbide precipitation is MC-type carbide → 0 → in terms of the component ratio in atom, according to EDS analysis. Then, the formation of the γ/γ0 eutectic phases and the γ0 phase simultaneously occurred during the solidification. As the solidification proceeds, the formation of the η phase represents the ending process of the GTD222 superalloy solidification. Owing to the alloying elements redistribution in the solidification process, the temperature of the solidus of the GTD222 superalloy, 1310 ◦C, is slightly lower than the liquidus temperature, which is 1360 ◦C. 2. As the cooling rate increased from 2.5 K/min to 20 K/min, the dendrite arm spacing decreased from 200 µm at 2.5 K/min to 100 µm at 20 K/min. Also, the size of the precipitation phase decreased as the cooling rate increased. According to these three precipitation phases, the onset temperature of precipitation phases varied as the cooling rate of the alloy changed.

Author Contributions: Conceptualization, Y.S. (Yanle Sun) and R.W.; methodology, B.G. and R.W.; softeware and validation, B.G., R.W. and Y.S. (Yanfei Sui); formal analysis, R.W. and Y.S. (Yanle Sun); Investigation, resources and date curation, B.G. and R.W.; writing—original draft preparation, R.W.; writing—review and editing, Y.S. (Yanle Sun); visualization, supervision, project administration and funding acquisition, H.W., C.Z. and Z.W. Funding: This research received no external funding. Conﬂicts of Interest: All the authors, Bo Gao, Yanfei Sui, Hongwei Wang, Chunming Zou, Zunjie Wei, Rui Wang and Yanle Sun, declare that no conﬂict of interest exits in the submission of this publication.

References

1. Panwisawas, C.; Mathur, H.; Gebelin, J.-C.; Putman, D.; Rae, C.M.; Reed, R.C. Prediction of recrystallization in investment cast single-crystal superalloys. Acta Mater. 2013, 61, 51–66. [CrossRef] 2. Kun, B.; Rui, H.; Jinshan, L.; Lian, Z. Grain refinement of Ni-Cr-W based superalloy by near liquidus casting. Rare Met. Mater. Eng. 2014, 43, 1–5. [CrossRef] 3. Chapman, L. Application of high temperature DSC technique to nickel based superalloys. J. Mater. Sci. 2004, 39, 7229–7236. [CrossRef] 4. Sun, C.; Kirk, M.; Li, M.; Hattar, K.; Wang, Y.; Anderoglu, O.; Valdez, J.; Uberuaga, B.; Dickerson, R.; Maloy, S.; et al. Microstructure, chemistry and mechanical properties of Ni-based superalloy Rene N4 under irradiation at room temperature. Acta Mater. 2015, 95, 357–365. [CrossRef] 5. Zupaniˇc,F.; Bonˇcina,T.; Krizman, A.; Tichelaar, F. Structure of continuously cast Ni-based superalloy Inconel 713C. J. Alloy. Compd. 2001, 329, 290–297. [CrossRef] 6. Chiou, M.-S.; Jian, S.-R.; Yeh, A.-C.; Kuo, C.-M.; Juang, J.-Y. High temperature creep properties of directionally solidified CM-247LC Ni-based superalloy. Mater. Sci. Eng. A 2016, 655, 237–243. [CrossRef] 7. Kontis, P.; Yusof, H.M.; Pedrazzini, S.; Danaie, M.; Moore, K.; Bagot, P.; Moody, M.; Grovenor, C.; Reed, R.; Bagot, P. On the effect of boron on grain boundary character in a new polycrystalline superalloy. Acta Mater. 2016, 103, 688–699. [CrossRef] Materials 2019, 12, 1920 9 of 9

8. Chen, Y.; Françis, E.; Robson, J.; Preuss, M.; Haigh, S. Compositional variations for small-scale gamma prime

(γ0) precipitates formed at diﬀerent cooling rates in an advanced Ni-based superalloy. Acta Mater. 2015, 85, 199–206. [CrossRef] 9. Zheng, L.; Gu, C.; Zheng, Y. Investigation of the solidiﬁcation behavior of a new Ru-containing cast Ni-base superalloy with high W content. Scr. Mater. 2004, 50, 435–439. [CrossRef]

10. Yang, W.; Liu, G.; Wu, K.; Hu, B. Influence of sub-solvus solution heat treatment on γ0 morphological instability in a new Ni-Cr-Co-based powder metallurgy superalloy. J. Alloy. Compd. 2014, 582, 515–521. [CrossRef] 11. Zhang, Y.; Huang, Y.; Yang, L.; Li, J. Evolution of microstructures at a wide range of solidification cooling rate in a Ni-based superalloy. J. Alloy. Compd. 2013, 570, 70–75. [CrossRef] 12. Li, X.; Wang, L.; Dong, J.; Lou, L. Effect of solidification condition and carbon content on the morphology of MC carbide in directionally solidified nickel-base superalloys. J. Mater. Sci. Technol. 2014, 30, 1296–1300. [CrossRef] 13. Seaver, D.; Beltran, A. Nickel base alloy GTD-222, a new gas turbine nozzle alloy. In Proceedings of the ASME International Gas Turbine and Aeroengine Congress and Exposition, Orlando, FL, USA, 3–6 June 1991. 14. Banik, T.; Deragon, T.C.; Schweizer, F.A. The effects of ingot composition and conversion on the mechanical properties and microstructural response of GTD-222. Superalloys 1996, 607–611. [CrossRef] 15. Kavoosi, V.; Abbasi, S.; Mirsaed, S.G.; Mostafaei, M. Influence of cooling rate on the solidification behavior and microstructure of IN738LC superalloy. J. Alloy. Compd. 2016, 680, 291–300. [CrossRef] 16. Sharma, G.; Chatterjee, A.; Paul, B.; Varshney, J.; Chakravartty, J. High temperature mechanical behavior of nanocrystalline Ni. Mater. Sci. Eng. A 2013, 576, 69–73. [CrossRef] 17. Schneibel, J.; Liu, C.T.; Miller, M.; Mills, M.; Sarosi, P.; Heilmaier, M.; Sturm, D. Ultrafine-grained nanocluster-strengthened alloys with unusually high creep strength. Scr. Mater. 2009, 61, 793–796. [CrossRef] 18. Zhai, W.; Wang, W.; Geng, D.; Wei, B. A DSC analysis of thermodynamic properties and solidification characteristics for binary Cu–Sn alloys. Acta Mater. 2012, 60, 6518–6527. [CrossRef] 19. Wang, C.; Guo, Y.; Guo, J.; Zhou, L. Investigation and improvement on structural stability and stress rupture properties of a Ni–Fe based alloy. Mater. Des. 2015, 88, 790–798. [CrossRef] 20. Antonov, S.; Huo, J.; Feng, Q.; Isheim, D.; Seidman, D.N.; Helmink, R.C.; Sun, E.; Tin, S. σ and η Phase formation in advanced polycrystalline Ni-base superalloys. Mater. Sci. Eng. A 2017, 687, 232–240. [CrossRef] 21. Chang, L.; Jin, H.; Sun, W. Solidification behavior of Ni-base superalloy Udimet 720Li. J. Alloy. Compd. 2015, 653, 266–270. [CrossRef] 22. Mao, J.; Chang, K.-M.; Yang, W.; Ray, K.; Vaze, S.P.; Ferrer, D.U. Cooling precipitation and strengthening study in powder metallurgy superalloy U720LI. Met. Mater. Trans. A 2001, 32, 2441–2452. [CrossRef] 23. Sun, W.; Lee, J.; Seo, S.; Choe, S.; Hu, Z. The eutectic characteristic of MC-type carbide precipitation in a DS nickel-base superalloy. Mater. Sci. Eng. A 1999, 271, 143–149. [CrossRef] 24. Seo, S.; Kim, I.; Lee, J. Eta phase and boride formation in directionally solidified Ni-base superalloy IN792 + Hf. Metall. Trans. A 2007, 38, 883–893. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).